97% model accuracy. 100 hours of training. No holding back.
Make an account and download on the Play Store Get instant access to our AI-powered plastic scanner.
Click the scan page to take a picture of a plastic item. Max 1 scan will be saved per 24 hours.
Once you've scanned enough plastic items, browse through an awesome catalog of prizes!
Extracting problem statement...
Here is a quick code sample that shows how EcoPlastic uses AI image classifcation to identify 100s of plastic items with 97% accuracy
1import tensorflow as tf
2import numpy as np
3
4# Data: Plastic recycled (X) vs CO2 reduced (Y)
5X = np.array([5, 10, 15, 20, 25], dtype=float) # Tons of plastic recycled
6Y = np.array([50, 100, 150, 200, 250], dtype=float) # CO2 reduced in kg
7
8# Define a simple model
9model = tf.keras.Sequential([
10 tf.keras.layers.Dense(units=1, input_shape=[1]) # One input, one output
11])
12
13# Compile and train the model
14model.compile(optimizer='adam', loss='mean_squared_error')
15model.fit(X, Y, epochs=100, verbose=0)
16
17# Convert the model to TensorFlow Lite format
18converter = tf.lite.TFLiteConverter.from_keras_model(model)
19tflite_model = converter.convert()
20
21# Save the TFLite model to a file
22with open('eco_plastic_model.tflite', 'wb') as f:
23 f.write(tflite_model)
24
25print("Model converted to TFLite and saved.")
26This is a quick summary of all the cool stuff EcoPlastic AI can do!.
Everything you need to know about EcoPlastic AI
Download and use EcoPlastic AI today.